`
`INET’95 Conference
`Proceedings
`
`Editor: Kilnam Chon
`
`International Networking Conference
`Honolulu, Hawaii, USA
`27-30 June 1995
`
`The Annual Conference of the Internet Society
`
`To order copies, contact the ISOC Secretariat at:
`Internet Society
`12020 Sunrise Valley Dr., Suite 270
`Reston, VA 22091
`USA
`isoc@isco.org
`
`RPX Exhibit 1222
`RPX v. DAE
`
`
`
`Reliable Audio for Use over the Internet
`
`Proc. INET ’95
`
`V.Hardman, A.Sasse. M.Handley, A.Watson
`
`Reliable Audio for Use over the Internet
`Vicky Hardman <v.hardman@cs.ucl.ac.uk>, Martina Angela Sasse <a.sasse@cs.ucLac.nk>,
`Mark Handley <m.handley@cs.ucl.ac.uk>, Anna Watson <a.watson@cs.ucl.ac.uk>
`
`Abstract
`This paper describes current problems found with
`audio applications over the MBONE (Multicast Back-
`bone). and investigates possible solutions to the most
`common one - packet loss. The principles of packet
`speech systems are discussed, and how the structure
`allows the use of redundancy to design viable solu-
`tions to the problem. The paper proposes the use of
`synthetic speech coding algorithms (vocoders) to pro-
`vide redundancy, since the algorithms produce a very
`low bit-rate stream, which only adds a small over-
`head to a packet. Preliminary experiments show that
`normal speech repaired with synthetic quality speech
`is intelligible, even at very high loss rates.
`
`Introduction
`The application of this work is multimedia confer-
`encing over the MBONE (Multicast Backbone), an
`experimental overlay network of the Internet. The
`work has arisen from experiences in multi-way multi-
`media conferencing in Project MICE (Multimedia In-
`tegrated Conferencing for Europe) [1], is currently
`applied in Project ReLaTe (Remote Language Teach-
`ing over SuperJANET) [2], and includes formal ex-
`periments into the human perception of packet speech
`systems degraded by packet loss.
`If multimedia conferencing is to become widely
`used in the Internet community, user must perceive
`the quality to be sufficiently good for most applica-
`tions. Experience has shown that audio is almost al-
`ways the most important component of multimedia
`conferencing. Whilst we have identified a number of
`problems which impair the quality of audio, the ma-
`jor one with audio over the MBONE is packet loss
`[3]. This paper attempts addresses the problem of
`packet loss over the MBONE.
`Packet loss can occur for a number of reasons:
`- congestion of routers and gateways, which lead
`to packet being discarded;
`- delays in packet transmission, with packet arriv-
`ing too late at the receiver to be played back;
`- heavy loading of the workstations, leading to-
`scheduling difficulties in multi-tasking operat-
`ing systems.
`Packet loss is a persistent problem, particularly
`given the increasing popularity, and therefore in-
`creasing lead, of the Internet. Possible ways of com-
`batting congestion include bandwidth reservation and
`moves toward an integrated service management on
`the Internet. These would require wide-scale changed
`
`to be agreed and implemented, so these solutions will
`be available in the short to medium term. Yet, the
`disruption of speech intelligibility even atJow loss
`rates which we currently experience may convince a
`whole generation of users that multimedia conferenc-
`ing over the Internet is not viable. We therefore pro-
`pose a solution which renders the speech intelligible
`under current network conditions, and can be de-
`ployed in the short term. Such a solution will have to
`be at the application level, i.e. the multicast audio
`tools.
`Current audio applications repair lost packets with
`silence, which leads to the speech clipping effects
`currently experienced by many users. Since compara-
`tively large packets are used, even the loss of individ-
`ual packet loss has a serious impact on the
`intelligibility of speech.
`We propose a method of repairing damaged
`speech using cheap redundancy within the packets
`sent from the transmitter. The redundancy is synthetic
`speech, which, when split into packets, only adds a
`very small amount of overhead, and therefore does
`not add to the congestion at the network level. The re-
`dundancy for any given packet of speech is piggy-
`backed onto a later packet. This mechanism means
`that when the receiver suffers the loss of the primary
`speech information, it still has the possibility of sub-
`stituting something sensible in the output stream of
`speech, provided that the redundancy can be received.
`In order to establish the effectiveness of this solu-
`tion, we have performed experiments into user per-
`ception of speech repaired with a synthetic substitute,
`the experiments subjectively measured speech intel-
`ligibility, and the results show that this technique is
`very successful at repairing speech with large packet
`sizes and for very high loss rates (results were taken
`up to 40%). The paper also describes how the pro-
`posed solution scales in the multicast environment.
`
`Background
`Speech Coding
`Speech coding schemes have been standardised
`for use pver telephone networks; a variety of speech
`coding algorithms exist for a single target quality of
`service (QoS), and at a very few discrete bit-rates:
`Pulse Code Modulation (PCM) operates at 64 kbps.
`Adaptive Differential Pulse Code Modulation (AD-
`PCM) operates at 32 kbps, and Code Excited Linear
`Prediction (LD-CELP) operates at 16 kbps. The target
`QoS is ’toll’ (or telephone) quality, and each algo-
`rithm available at this QoS produces a different bit-
`rate: the improvement in bit-rate being obtained for
`
`This material may be protected by Copyright law (Title 17 U.S. Code)
`
`171
`
`
`
`Reliable Audio for Use over the Internet
`
`Proc. INET ’95
`
`V.Hardman, A.Sasse, M.Handley, A.Watson
`
`increasing complexity in the coding algorithms. An-
`other standard speech coding algorithm is Groupe
`Speciale Mobile (GSM), which was designed for use
`over cellular telephone networks. The target QoS is
`consequently slightly less than toll, but the algorithm
`is popular, since it operates at the same bit-rate as
`CELP, but is much less complex. A fuller discussion
`of toll quality speech coding algorithms can be found
`in [4].
`A second class of coding algorithms exist, which
`operate at the ’communications’ or synthetic QoS.
`These algorithms operate at very low bit-rates (ap-
`prox. 4.8kbps and below), and produce very mechan-
`ical sounding speech. Perhaps the most important
`method of this class is Linear Predictive Coding
`(LPC), since the principle is also an integral part of
`both the CELP and GSM coders. A fuller description
`of which can be found in [5].
`Packet Speech Systems
`Packet speech systems usually employ the stand-
`ard speech coding algorithms, and group the emerg-
`ing stream of codewords into packets for transmission
`over the network. At the receiver, the packets may be
`delivered: out of order, not at all, or at non-uniform
`intervals. Consequently, a reconstruction delay must
`be used at the receiver to repair the network effects;
`this enables sample play-out to be smoothed.
`In a packet speech system, the end-to-end delay is
`always a critical factor in the usability of a real-time
`voice system, and should be kept below 600ms in the
`absence of echoes (The figure may be in fact be less
`than this - 400ms) [6], if conversation patterns are not
`to break down. The size of fhe packets (in ms) chosen
`for a packet speech-system directly impacts the end-
`to-end delay. A delay equal to the size of one packet
`is incurred at the transmitter, since the samples in the
`packet have to be collected before a packet can be
`sent. At the receiver, a rough estimate of the recon-
`struction delay required to smooth out packet arrival
`times is two packets worth in ms [7] [8], although the
`true value may be substantially in excess of this rule
`of thumb. Consequently, a minimum of three packets
`worth of delay is incurred on an end-to-end basis, be-
`fore the network propagation delay has been taken
`into account.
`The delay introduced will be enough to receive
`most of the packets, but some will always arrive too
`late to be played back, and can be considered ’lost’.
`Furthermore, the network itself may lose packets. In
`such situations, the speech ’stream’ must be repaired,
`and a dummy packet inserted in place of the lost one,
`so that the correct timing relationship is maintained
`between the transmitter and receiver. The presence of
`the dummy packet is usually discernible to the listen-
`er, and unfortunately, the perceptibility of the loss in-
`creases with increasing packet size, as well as with
`increasing loss rate.
`The impact of the two factors identified above,
`(delay and loss), is such that small packet sizes are re-
`quired for real-time voice links. However, the use of
`
`small packets increases the overhead of packet head-
`ers, and any processing incurred at network nodes, ^
`and therefore increases the likelihood of congestion
`and loss. Consequently, a trade-off exists between the
`requirements of the network, and the requirements of
`real-time voice connections.
`Voice Reconstruction Techniques
`Repair methods for packet loss are known as voice
`reconstruction mechanisms. The aim is to construct a
`suitable dummy packet at the receiver, so that the loss
`is as imperceptible as possible. With compressed
`speech, voice reconstruction mechanisms not only
`have to produce a suitable fill-in packet, but also have
`to maintain the decoder tracking, since the algorithms
`transmit difference information. Voice reconstruction
`techniques can be split into two categories; receiver
`only, and combined source and channel techniques.
`
`Table 1: Voice Reconstruction
`Techniques
`
`Combined Source and
`Channel
`
`Embedded Coding
`
`Redundancy
`
`Receiver-Only
`
`Silence
`
`White Noise
`
`Waveform Substitu-
`tion
`
`Sample Interpolation
`
`Receiver-only techniques are those that try to re-
`construct the missing segment of speech solely at the
`receiver, possibly from correctly received packets
`preceding that which was lost. Combined source and
`channel techniques are those that try to make the sys-
`tem robust to loss by either arranging for the transmit-
`ter to code the speech in such a way as to be robust to
`packet loss, or by transmitting extra information to
`help with reconstruction.
`Receiver-OnlyT echniques
`The original voice reconstruction techniques were
`receiver-only, and used either silence, white noise,
`repetition of part of the last correctly received speech
`waveform, or sample interpolation as the substitute.
`Silence substitution is favoured because it is sim-
`ple to implement, and it gives adequate performance
`for small packet sizes (<16ms), and up to 1% loss [9]
`[10].
`It is well known that other methods give substan-
`tially better results than those obtained from silence
`substitution. Warren [11] investigated the human per-
`ception of speech interrupted by silence compared to
`noises, such as coughs. The results show that phone-
`mic restoration (the ability of the human brain to sub-
`consciously repair the missing segment of speech
`with the correct sound) occurs for the noise situation.
`
`172
`
`
`
`Reliable Audio for Use over the Internet
`
`Proc. INET ’95i
`
`V.Hardman, A.Sasse, M.Handley, A.Watson
`
`and does not occur for silence substitution.
`Experience from the MICE project has shown that
`listeners can become very frustrated with MBONE
`speech. Their frustration stems from^a variety of au-
`dio problems:
`- interrupted speech due to packet loss - the speech
`sounds at first bubbly, and then individual in-
`terruptions are apparent;
`- talking into a ’dead’ channel, and the lack of au-
`tomatic gain control for listeners;
`- high levels of background noise cutting in and
`out;
`f
`- distortion due to overloading the microphone
`when speakers shout, or mis-matched levels.
`Packet loss is the most ^strating problem, and
`one users cannot cure of their own accord. The fhis-
`tration with interrupted speech can be explained by
`considering the linguistic construct of a sentence,
`which includes a pause (of duration > a phoneme) at
`the end of the sentence. Since the size of packets used
`over the MBONE are often comparable to the length
`of a phoneme, the interruptions in the speech flow
`sometimes occur at inappropriate points, which sends
`ambiguous signals to brain, as to whether speech is
`continuing or not [11].
`White noise was shown to give a subjective per-
`formance improvement over silence by Miller [12]
`when contextual information in speech was removed,
`and an intelligibility improvement by Warren [11]
`when the contextual information was present. Conse-
`quently, silence substitution is not a suitable means of
`voice reconstruction, since white noise is known to
`give improvements, and is as easy to generate as si-
`lence.
`Other receiver-only voice reconstruction tech-
`niques rely on the assumption that the speech charac-
`teristics have not changed from a precfeding segment
`of speech, and use this preceding segment informa-
`tion to reconstruct the missing part; a simple example
`of this sort of voice reconstruction would be to repeat
`the last correctly received packet.The mechanisms
`fail when the packet sizes are large, and the loss rate
`is high (packets are more likely to be lost in twos or
`threes, than singularly). A fuller explanation of exist-
`ing receiver-only techniques can be found in [13].
`Combined Source and Channel'Techniques
`Combined source and channel techniques gener-
`ally show significant improvement over receiver only
`techniques. The techniques either transmit extra in-
`formation within the speech packets (to help with re-
`construction at the receiver), or alter the speech
`coding algorithm and network operation (to make
`system as a whole more robust to packet loss).
`Embedded speech coding techniques used with
`adaptive differential pulse code modulation (ADPCM
`[14]), such as those by [15], [16], and code excited
`linear prediction (CELP) [17], have shown significant
`
`performance improvements during packet loss.’Em-
`bedded speech coding techniques allow the bit-rate
`can be adjusted from 40 to 32 or 23 kbps, without the
`introduction of large amounts of noise; essentially the
`feed-back loops in the encoder and decoder operate at
`a lower resolution than usual. The standard was de-
`signed to ease the problem of packet loss in packet
`networks; the codewords are segmented into high and
`low priority bits, and then placed in different packets.
`The mechanism relies on arranging for the network to
`drop packets containing LSBs only, which means that
`the mechanism is not applicable to networks which do
`not provide this support, such’ as today’s Internet.
`Lara-Barron [15] investigated embedded AD-
`PCM coding techniques at 16-32 ms, and reported
`success for up to 40% loss (no reduction in speech
`quality for up to 6% loss).
`The significant improvement resulting from the
`use of this mechanism is mostly due to the preserva-
`tion of the decoder adaption logic [16].
`Speech Quality
`Speech quality may be assessed by either subjec-
`tive or objective means, although it is well known that
`subjective assessment methods provide more accu-
`rate results.
`Subjective assessment is usually made by per-
`forming listening tests using a large number of sub-
`jects.'The material used, and the measurements made,
`depend upon the likely degree of distortion expected.
`Toll quality speech coding algorithms are usually
`assessed by mean opinion scores (MOS) [18], where
`encoding distortion and noise are the likely type of
`degradation suffered. The technique involves the lis-
`tener making a category rating after listening to a pas-
`sage of speech.
`Synthetic quality speech coding algorithms result
`in speech that has far greater degradation than found
`in toll quality systems; intelligibility .is usually only
`adequate at best. Consequently, the MOS method is
`not suitable, and communications quality systems are
`assessed using comprehension or intelligibility tests.
`There is a wide range of speech material available,
`ranging from a sequence of syllables '(the listeners
`transcribe what they hear) to passages (the compre-
`hension of which is ascertained by asking a series of
`questions) [19].
`The speech material is chosen based on the re-
`quired sensitivity of the results, desired experimental
`control, and range of human faculties included in the
`test.
`Reliable Audio for Use over the Internet
`We have developed a new voice reconstruction
`scheme that uses redundancy to improve voice recon-
`struction at the receiver.
`The redundant information is the output of a syn-
`thetic quality speech coding algorithm (LPC), which
`
`
`
`Reliable Audio for Use over the Internet
`
`Proc. INET ’95
`
`V.Hardman, A.Sasse, M.Handley, A.Watson
`
`is very low bit-rate (4.8kbps). LPC is generally con-
`sidered to contain about 60% of the information con-
`tent of the speech signal, as the overall shape of the
`frequency spectrum is preserved at the expense of
`short-term amplitude and pitch variations. This tech-
`nique is exactly what is required for successful voice
`reconstruction; the gap will be filled with a sound that
`is expected, and phonemic restoration should im-
`prove the situation further.
`The Characteristics of the MBONE / Internet
`The Interiiet, and its multicast overlay (MBONE)
`is a unique ’shared’ packet network, that offers scala-
`ble multi-way communication. Such a network has
`traditionally not been considered suitable for speech
`applications, because of the large end-to-end delay
`that is commonly experienced ovpr the network, and
`the potentially high probability of packet loss, (with
`relatively large segments of speech being lost). Cur-
`rent audio tools such as vat [20], and nevot [21] have|
`however, demonstrated that these problems are not
`prohibitive to successful voice communications,
`since use of these tools is very widespread.
`The Internet provides variable length packets, a
`feature which has the potential for fine-grained con-
`trol over the trade-off between network and speech
`performance requirements. The ’per-packet’, rather
`than ’size-.of-packet’ network penalty for small pack-
`ets coupled with the ability to have variable length
`packets also means that the state information from the
`speech coding algorithms can be transmitted in the
`packet, which substantially improves the perception
`of packet .loss. Current audio tools transmit coding al-
`gorithm st^te information in each packet, but replace
`lost packets with silence.
`The Loss Charaeteristics of the MBONE
`Current research by a MICE partner, Bolot [22], is
`investigating the numljer of consecutive losses found
`over the MBONE. The results show that for light and
`intermediate loads, losses are essentially non-consec-
`utive for an audio stream, and for heavy loads, the be-
`haviour is similar, but consecutive losses are more
`prevalent.
`These results suggest that a model where the re-
`dundancy is positioned in the packet after speech
`from the primary coding algorithm is suitable for light
`and medium network loads, and a model with the re-
`dundancy positioned a number of packets later is suit-
`able for heavy loads.
`Voice Reconstruction for the MBONE
`LPC as the redundant information adds only a
`small amount of overhead to an RTP [23] packet (12
`bytes per 160 bytes of PCM (or per 80 bytes of AD-
`PCM). The information is piggy-backed to the packet
`following that containing the primap^ speech code-
`words; that the loss of an individual jiacket can be re-
`paired using the redundant information in the
`following packet. This mechanism is unique to packet
`networks, and is only feasible because of the recon-
`
`174
`
`struction delay introduced at the receiver.
`The use of this redundancy technique means an in-
`crease in the reconstruction delay by the time equiva-
`lent of the distance of the redundancy component
`after the primary component; this implies an extra de-
`lay of one packet for light and medium loading con-
`ditions.
`Multiple multicast receivers in a single conference
`may experience a variety of the characteristics report-
`ed in [22]. Consequently, the reconstruction mecha-
`nism may occasionally have more than one instance
`of the redundancy after the primary coding scheme
`packet. In this way, the heavy loading characteristics
`seen by one site do not affect the performance of the
`majority.
`
`Redundancy
`Primary Speech
`Coding
`
`a)
`
`b)
`
`X 1
`4 1
`;V
`6 1
`
`m
`5 >
`m
`
`3 E
`
`' 4 I
`
`V
`
`2 1
`
`3 i
`
`Figure 1: The Positioning of Redundancy
`Relevant to the Primary Coding for
`a) low and b) higher loss rates
`
`The provision of LPC redundant information for
`use in voice reconstruction is intended to be used with
`per-packet state information; this prevents decoder
`mistracking in the case of loss. When a packet has
`been lost, the receiver decodes the redundant infor-
`mation, and feeds the samples to the audio hardware.
`Consequently, the output speech waveform consists
`of periods of toll quality speech, interspersed with pe-
`riods of synthetic quality speech.
`While LPC is a fairly complex speech coding al-
`gorithm, it should be noted that linear predictive anal-
`ysis and synthesis are an essential part of all new
`higher compression schemes: GSM and CELP both
`use these techniques as a first step in their algorithms.
`LPC also has the potential to be used elsewhere in the
`system; as an improvement to the silence detection
`function, which is an integral part of most packet
`speech systems.
`Experimental Design
`Voice reconstruction experiments to date have
`usually been performed with packet sizes of 16-32ms,
`or less. The packet sizes used over the MBONE are
`usually greater than these values (40ms is commonly
`used). Consequently, little information exists about
`the degradation commonly experienced in voice con-
`
`
`
`Reliable Audio for Use over the Internet
`
`Proc. INET ’95
`
`V.Hardman, A.Sasse, M.Handley, A.Watson
`
`A detailed description of the experimental details
`can be found in [13] and [26].
`
`Results
`The interaction between reconstruction scheme
`and packet size is difficult to analyse, since other fac-
`tors, such as the temporal characteristics of speech
`sounds, and the masking and temporal order percep-
`tion capabilities of the ear also affect human percep-
`tion.
`
`Silence Substitution
`The first point to consider is whether the results
`from these experiments give an insight into the
`speech perception from current audio tools.
`______Silence
`___Repetition
`___LPC Repair
`%Articulation
`nc____________ -
`yj
`-
`_______________ I________________
`1 -- -- •_ -- -- t A DPPM
`•S.
`“l-xj ... )............. No Loss
`
`t. —
`\
`
`-1
`
`nections over the MBONE/Intemet. This experiment
`was therefore designed to compare LPC redundancy
`with waveform substitution dfld silence substitution;
`receiver-only techniques, such as-waveform substitu-
`tion are cheap to implement, and could potentially be
`used instead of LPC under certain circumstances. The
`experiment was also designed to try and give a broad
`outlook on the question of the degradation experi-
`enced as a result of packet loss over the Internet, by
`considering a wide range of loss rates (0-40%).
`Waveform substitution was chosen as a represent-
`ative receiver-only method, and the simplest of these,
`packet repetition, was considered suitable for com-
`parison.
`It was assumed that LPC redundancy could al-
`ways be received in the event of loss - this assumption
`does not hold true in the real world, but provides a
`valid basis for the experiment design.
`Waveform substitution repeats the last correctly
`received packet until the period of loss ends. This
`mechanism has an inherent draw-back when the lost
`packet is the last in a talk-spurt; the last packet will be
`repeated until a new talk-spurt starts.
`The solution to this problem lies in realising that
`there should be a limit on the length of ’waveform
`substituted’ speech, before the underlying'assumption
`of no change in the speech characteristics breaks
`down; a suitable figure for this is 80ms, or the average
`length of a phoneme. Consequently, when the packet
`size is 20ms, the last correctly received packet may be
`repeated three times. When the packet size is 40 ms,
`the packet may be repeated twice (120ms). When the
`packet size is 80ms, the packet may be repeated once
`(160ms).
`The speech was recorded and manipulated using a
`Sun SPARC station 10, the OGI speech tools soft-
`ware [24], which is a system that was developed for
`speech recognition experiments, and software written
`by the author (to generate the loss, code and decode
`the speech using different algorithms etc.). The co-
`deds used in these experiments are publicly available
`versions [ADPCM][LPC], with ADPCM conforming
`to the CCITT standard [25]. The loss was generated
`randomly.
`The subjective quality was assessed using phonet-
`ically balanced (PB) words, which proportionally
`represent the sounds found in every-day English
`speech.
`There were three groups of seven subjects, those
`who heard the PB word lists reconstructed by silence
`substitution, those who heard waveform substitution,
`and those who heard LPC reconstruction. Each sub-
`ject in each group had ten different lists, nine of which
`were test conditions, and the first of which was a no
`loss control condition. The lists were 25 words long,
`and after recording answers, for each list, the subjects
`completed a MOS rating scale, rating the quality of
`the speech just heard on a five-point scale, from bad
`to excellent.
`
`A..'. 7
`\ :
`\:
`
`'
`
`•.
`
`'
`
`10
`
`20
`
`30
`
` LPC
`40'"'“"
`%Loss
`lire 2: Voice ReconstructiQn for 20/40ms Packets
`
`85
`
`75
`
`65
`
`55
`
`45
`
`35
`0
`
`Fig
`
`Referring to figure 2: Silence Substitution for 20
`and 40 ms packet sizes, it can be seen that the graph
`indicates that using silence substitution for the voice
`reconstruction scheme fails between 15 and 20% loss.
`For 80ms packets (see figure 3), the results suggest
`that intelligibility is inadequate at a 15% loss rate.
`This observation is consistent with experiences
`obtained from Project MICE. The observation is also
`in keeping with findings reported by [12] and [27];
`’speech degradation as high as 50% can be tolerated
`(intelligibility of 80%) if the packet size is small
`(0.019s)’ on the other hand, if packets are long
`(0.25s), and the loss probability is high, intelligibility
`decreases to very low values (10%)’.
`Waveform Substitution
`The results for waveform substitution indicate
`that, for a packet size of 20ms, intelligibility does not
`decrease significantly with loss rate over the range of
`measurements taken.’ When the packet size is 40 ms,
`however, a significant difference is found between
`30% loss and 40% loss, whilst with 80 ms packets sig-
`nificant differences can be found at much lower levels
`of loss: performance drops between 15 and 20%, and
`between 20 and 30%.
`
`
`
`Reliable Audio for Use over the Internet
`
`Proc. INET ’95
`
`V.Hardman, A.Sasse, M.Handley, A.Watson
`
`Conclusions and Further Work
`This paper has described a voice reconstruction
`method for use in packet networks. The mechanism is
`suitable for both unicast and multicast connections
`under all types of network conditions, although the
`main intelligibility advantage is to be found with
`large packet sizes, and medium to high loss rates; a re-
`ceiver only voice reconstmction mechanism is suita-
`ble for light loss rates and small packet sizes.
`Listeners however, prefer LPC reconstruction for
`packet sizes of 40 and 80ms.
`The mechanism may have minimal overhead in
`processor utilisation, since LPC analysis is part of the
`more complex coding algorithms, and may be used in
`the future to enhance other aspects of an audio tool.
`The method also only adds a small amount of over-
`head to the bandwidth used in a voice conference.
`Since the overhead in terms of bandwidth is small,
`it may be desirable to have multiple copies of the re-
`dundancy present. This rationale would enable an au-
`dio tool to provide multi-cast international
`conferences with a voice reconstruction mechanism
`that scales; receivers at the end of ’good’ network
`branches would have good performance with imper- •
`ceptible loss, .while receivers at the end of ’poor’ net-
`work branches would experience a small reduction in
`speech quality, but would have a larger end-to-end
`delay.
`The paper describes formal speech quality tests on
`the human perception of speech at different packet
`loss rates. The results show that the perceptibility of
`packet loss need no longer be regarded as one of the
`main constraints on the packet size if LPC redundan-
`cy is used to aid voice reconstruction.
`In particular, the voice reconstruction should take
`the following form:
`At low loss rates (20% and below) and small
`packet sizes (20 and 40ms) the results show that a re-
`ceiver only technique is suitable, although LPC re-
`dundancy should also be used for 40ms packets, since
`listeners preferred it to waveform substitution.
`At higher loss rates and for all conditions when
`packet size is 80ms, LPC redundancy should be used.
`The MICE project is currently implementing a
`prototype audio tool, which will include voice recon-
`struction using LPC redundancy. Techniques to pro-
`vide controlled feed-back from multiple receivers in
`multi-cast conferences are also being developed,
`which will enable the use of redundancy to be adap-
`tive to receiver’s needs. Further studies on human
`perception of audio and network performance, are
`planned, which will investigate the limits on packet
`size; the study will include an analysis of the impact
`of delay, as well as the human perception of large
`packet loss. The human perception studies will also
`address the relationship between objective measures
`(such as network performance results), and subjective
`perception and performance.
`
`Comparison of Silence Substitution and
`Waveform Substitution
`A comparison between the two receiver-only
`voice reconstruction schemes was made for different
`packet sizes. The results show that waveform substi-
`tution is better than silence substitution for packet siz-
`es of 20 and 40ms, but that the advantage is not
`present for packet sizes of 80ms. The highest advan-
`tage was obtained for packet sizes of 20ms.
`The reduction in advantage of voice reconstruc-
`tion scheme for 40ms (compared to 20ms) might be
`due to the practice within the voice reconstruction
`scheme of allowing a 40ms packet to be repeated up
`to twice, which implies an assumption that speech
`characteristics have not changed for 80ms. This as-
`sumption is unlikely to be valid all of the time, since
`the average length of a phoneme is 80ms. Better re-
`sults might be obtained by restricting repetition to one
`40ms interval only.
`LPC Redundancy
`Referring to Figure 2, it can be seen that for packet
`sizes of 20 and 40ms, as loss rate increases, intelligi-
`bility does not deteriorate much. Slight deterioration
`in intelligibility is only present for packet sizes of
`80ms, and then only at high loss rates.
`Comparison of LPC and Waveform Substi-
`tution
`For 20ms and 40ms packets, the results show that
`there is little advantage of using LPC instead of wave-
`form substitution.
`For 80ms packets, LPC should always be used.
`^ MOS Results
`The MOS results show the same trends as the in-
`telligibility results. However, the listeners preference
`for LPC redundancy instead of waveform substitution
`for 40ms packets was more marked than is shown in
`the intelligibility graph (Figure 2).
`
`176
`
`
`
`Reliable Audio for Use over the Internet
`
`Proc. INET ’95
`
`V.Hardman, A.Sasse, M.Handley, A.Watson
`
`Acknowledgements
`The ideas relating to using LPC for redundancy
`have been discussed at length in the technical meet-
`ings of the MICE (Multimedia Integrated Conferenc-
`ing for Europe - ESPRIT 7606 Project). In particular,
`we would like to acknowledge the collaboration with
`Christian Huitema and Jean Bolot from INRIA
`Sophia Antipolis, France, on this issue. Very special
`thanks are due to Van Jacobson (Lawrence Berkley
`Labs), the creator of vat, for many fruitful discus-
`sions and email exchanges on audio problems and
`how to cure them.
`
`References
`[1] Kirstein P.T., Sasse M.A., Handley M.J.,’Recent
`Activities in the MICE Conferencing Project’ Pa-
`per No. 166, INET 95.
`[2] Buckett J. Campbell I. Watson T.J. Sasse M.A.
`Hardman V.J. Watson A. ’ReLaTe: Remote Lan-
`guage Teaching over SuperJANET’ Proceedings
`of UKERNA 95, Networkshop, March 1995.
`[3] Sasse M. A. et al. ’Remote Seminars Through
`Multimedia Conferencing; Experiences from the
`MICE Project, Proceedings of INET94/JENC5,
`pp. 251-258.
`[4] Papamichalis P.E. ’Practical Approaches to
`speech Coding’ Publ. Prentice-Hall 1987.
`[5] Gold B. ’Digital Speech Networks’ Proceedings
`of the IEEE, Vol. 65, No. 12,